In the field of medical imaging, breast ultrasound has emerged as a crucial diagnostic tool for early detection of breast cancer. However, the accuracy of diagnosing the location of the affected area and the extent of the disease depends on the experience of the physician. In this paper, we propose a novel model called UGGNet, combining the power of the U-Net and VGG architectures to enhance the performance of breast ultrasound image analysis. The U-Net component of the model helps accurately segment the lesions, while the VGG component utilizes deep convolutional layers to extract features. The fusion of these two architectures in UGGNet aims to optimize both segmentation and feature representation, providing a comprehensive solution for accurate diagnosis in breast ultrasound images. Experimental results have demonstrated that the UGGNet model achieves a notable accuracy of 78.2\% on the "Breast Ultrasound Images Dataset."

The rapid growth of multimedia communication applications drives the demand for network communication, creating significant pressure on network systems and requiring effective queue management solutions to maintain performance and minimize congestion. With fluctuating traffic loads and increasing demands for Quality of Service, traditional queue management methods fail to meet the requirements. To resolve the challenge, the paper proposes the enhanced DAIM-RED model. DAIM-RED offers a comprehensive solution, leveraging adaptive techniques to adjust min-max thresholds, reduce packet drop probability, and optimize queue management efficiency in network routers. The model incorporates AIM-RED, an adaptive method capable of automatically updating the model and tuning parameters based on network data, along with Deep Q-Network, which predicts queue overflow conditions and optimizes throughput. DAIM-RED demonstrates superior network performance compared to models combining AIM-RED with Convolutional Neural Networks and Long Short-Term Memory. The model not only stabilizes queues but also minimizes congestion and ensures Quality of Service in increasingly complex network environments.

Biometric and identification patterns have gained extensive research and application, particularly in iris recognition. The iris harbors a wealth of individual-specific information, making it a vital element in biometric authentication. This article presents a comprehensive study encompassing iris segmentation and identification. We introduce the Weighted Adaptive Hough Ellipsopolar Transform Convolutional Neural Network (WAHET-CNN) as a novel approach for classifying pattern images. Our experimental outcomes demonstrate a commendable 90% accuracy achieved by the proposed WAHET-CNN on the CASIA dataset Version 4.

Facial emotion recognition technology is used to analyze and recognize human emotions based on facial expressions. This technology uses deep learning models to classify facial expressions, eyes, eyebrows, mouth, and other facial expressions to determine a person's emotions. The application of facial emotion recognition in the field of education is a potential way to evaluate the level of student absorption after each class period. Using cameras and emotion recognition technology, the system can record and analyze students' facial expressions during class. In this paper, we use the Convolutional Neural Network (CNN) algorithm combined with the linear regression analysis method to build a model to predict students' facial emotions over a period of time camera recorded.

Trong bối cảnh mạng hiện đại, tắc nghẽn lưu lượng và “bufferbloat” là những thách thức lớn, đòi hỏi các cơ chế điều khiển hàng đợi chủ động (AQM) tại bộ định tuyến để đảm bảo chất lượng dịch vụ. PIE (Proportional Integral Enhanced) là một thuật toán AQM hiệu quả, sử dụng bộ điều khiển PI để duy trì độ trễ mục tiêu, giảm thiểu tắc nghẽn và cải thiện hiệu suất mạng. PIE nổi bật nhờ tính đơn giản, ổn định và đã được triển khai rộng rãi trong các bộ định tuyến hiện đại. Tuy nhiên, với sự đa dạng của lưu lượng mạng thế hệ mới, PIE truyền thống gặp hạn chế trong việc tự động thích nghi với các tải động. Để khắc phục, bài báo đề xuất kết hợp PIE với Deep Q - Network (DQN) – một phương pháp Học tăng cường (RL). DQN giúp tối ưu hóa tham số PIE ( , , ) dựa trên dữ liệu mạng phức tạp, từ đó nâng cao khả năng phản ứng với các loại lưu lượng khác nhau. Kết quả thử nghiệm cho thấy, DQN-PI giảm độ trễ trung bình và tăng thông lượng so với PIE gốc, đặc biệt hiệu quả trong môi trường mạng không đồng nhất. Trong tương lai DQN - PI tiếp tục cải thiện hiệu suất AQM, thích ứng với nhiều môi trường mạng phức tạp như 5G, IoT, truyền phát thời gian thực hợp chuyển đổi số, nhằm củng cố lợi thế cạnh tranh lâu dài của địa phương.

Nowadays, the continuous development of information technology, communication over the Internet is increasing rapidly, and network congestion has become an alarming issue. To develop communication network infrastructure in a large city, a country, or globally, streamlining and controlling network data flow to optimize communication processes and minimize network congestion is crucial and necessary. In this study, the authors analyze and process data according to the delay of Internet Protocol (IP) packets, using machine learning models with the Random Forest (RF) and the Support Vector Machines (SVM) method to classify IP packets. The primary goal of classifying packets by delay is to optimize network performance by prioritizing processing of low-delay packets, ensuring stable and uninterrupted online services such as video streaming and voice calls. Furthermore, it is easy to manage and control packet traffic, hence minimizing network congestion at the router. ® 2024 Journal of Science and Technology - NTTU